Fig 1.
Example of en face angiograms and their binarization with the tested algorithms.
Columns and rows show methods and segmented plexuses, respectively.
Table 1.
Demographic and main clinical data of study population.
Fig 2.
Vessel density values calculated with the tested methods.
Vessel density values calculated with the tested methods for the entire cohort, diabetic retinopathy (DR) and glaucoma patients for the macular superficial capillary plexus (SCP) (A,B,C), deep capillary plexus (DCP) (D,E,F), and peripapillary radial capillary plexus (RCP) (G,H,I). Processing methods differences were significant at p < 0.0001 at all levels. SCP: superficial capillary plexus; DCP: deep capillary plexus; RCP: radial capillary plexus; DR: diabetic retinopathy.
Table 2.
Vessel density values estimated with the different methods.
Table 3.
Inter-algorithm agreement, reliability, and pairwise comparison in the superficial capillary plexus for the entire cohort of patients.
Table 4.
Inter-algorithm agreement, reliability, and pairwise comparison in the deep capillary plexus for the entire cohort of patients.
Table 5.
Inter-algorithm agreement, reliability, and pairwise comparison in the peripapillary radial capillary plexus for the entire cohort of patients.
Fig 3.
Inter-rater reliability of the manual method for the entire cohort, DR and glaucoma patients.
**** significant at p < 0.0001; *** significant at p < 0.001; ** significant at p < 0.01; NS: not significant; DR: diabetic retinopathy; ICC: intraclass correlation coefficient; op: operator; SCP: macular superficial capillary plexus; DCP: macular deep capillary plexus; RCP: peripapillary radial capillary plexus.
Fig 4.
Comparative plots for the seven processing methods for the superficial capillary plexus (SCP).
Green lines represent the no-bias line, while black circles demonstrate the true corresponding measurement among coupled devices.
Fig 5.
Comparative plots for the seven processing methods for the deep capillary plexus (DCP).
Green lines represent the no-bias line, while black circles show the true corresponding measurement among coupled devices.
Fig 6.
Comparative plots for the seven processing methods for the peripapillary radial capillary plexus (RCP).
Green lines represent the no-bias line, while black circles demonstrate the true corresponding measurement among coupled devices.
Table 6.
Calibration equations to convert vessel density value from one algorithm into the equivalent from another algorithm.
Fig 7.
Receiver operating characteristic (ROC) curves for discrimination of diseased from healthy eyes.
Legends show area under the ROC curve and 95% confidence interval. SCP: macular superficial capillary plexus; DCP: macular deep capillary plexus; RCP: peripapillary radial capillary plexus.